[Style]: Eastern Wuxia pinnacle duel, Extreme Kinetic Energy, aesthetics of environmental destruction, fast-paced editing, cinematic particle effects, epic score atmosphere. [Duration]: 15 seconds. [Characters]: Grandmaster in white VS Grandmaster in black. Shot 1: High-speed Collision. Both the white and black figures transform into afterimages, colliding violently in the center of an ancient stone platform. At the moment of weapon contact, blinding sparks and a visible ring-shaped high-pressure Shockwave burst out. The surrounding flagstone ground instantly cracks, blasting up debris and rubble. Shot 2: Mid-air Combat & Destruction. Both soar into the air, weaving and clashing at high speed amidst falling boulders, moving too fast for moves to be seen clearly. Every block triggers a visual sense of air distortion. They crash through a massive stone pillar; the pillar collapses and shatters into powder, dust filling the air, showcasing extreme destructive power. Shot 3: Ultimate Climax. Both land and simultaneously release their ultimate moves, causing two massive energy orbs (one cyan, one red) to clash against each other. This triggers an earth-shattering explosion of dust and energy. Mushroom-cloud-like smoke rises to obscure everything, and the entire base of the stone platform begins to collapse and sink.
[Style]: Eastern classical Xianxia aesthetics, aestheticism, cinematic quality, fluid cloth simulation, glowing particle effects, dreamy lighting. [Duration]: 15 seconds. [Protagonist]: Exquisite fairy, wearing multi-layered plain white ethereal feather-clothing, extra-long floating ribbons. Shot 1: Epic Wide Shot. Location is an ancient platform on a floating immortal island above a sea of clouds. The fairy stands still, backed by immense rays of morning glow (Tyndall effect). The focus is on the countless long silk ribbons on her body dancing extremely smoothly in the wind, spreading across the screen like water ripples, showcasing the ultimate elegance of floating fabric. Shot 2: Slow-motion Medium Shot. The fairy begins the classical Jinghong Dance (Frightened Swan Dance). She rises into the air (anti-gravity Qinggong feel), spinning and throwing out wide water sleeves. The lens focuses on the form of the silk sleeves blooming in the air and her elegant body language. Shot 3: Close-up & Interaction. Camera zooms in to a facial close-up. She lightly parts her red lips to begin singing (natural singing mouth shape). Accompanying the song, countless glowing pink flower petals appear out of thin air, forming a dazzling particle storm vortex around her, her eyes filled with divine radiance.
[Style]: Ultimate Girlfriend POV, Handheld Selfie VLOG, Vertical Aspect Ratio (9:16), Film Grain Filter, Natural Light (Golden Hour), camera includes natural shake and adjustments. [Duration]: 15 seconds. [Protagonist]: Taiwanese girl, long slightly curled hair, wearing a gentle knit cardigan, clean/sheer makeup, speaking tone is soft, slightly spoiled/coquettish with a complaining accent. [Scene]: Taipei Tamsui riverbank, background is the golden sunset, shimmering river surface, and the Lover's Bridge in the distance. Shot 1: The Rushed Intro. The camera shakes significantly because she is walking fast. She looks back at the camera (looking at you) while walking, brows slightly furrowed, pretending to be angry. One hand holds the phone, while the other hand holds down her hair which is being blown by the wind. Shot 2: Sharing the View. She stops walking, moves the camera away from her face, and turns it towards the stunning golden sunset across the river (Golden Hour). The camera displays the scenery for 2 seconds, then immediately turns back to focus on her face. Her face is illuminated golden by the sunset, showing a super happy smile. Shot 3: The Interaction. Suddenly, she is holding a super tall soft-serve ice cream unique to Tamsui. She first holds the ice cream up to the camera (as if feeding you), then quickly takes a lick herself. A bit of ice cream accidentally gets on the tip of her nose. She giggles foolishly at the camera.

无论你是 AI 视频创作的新手,还是希望提升技能的老手,这份 Seedance 3.0 提示词指南都能帮助你写出更好的提示词,释放 Seedance 3.0 AI 视频生成器的全部创作潜力。从基本的 Seedance 提示词结构到高级视频提示技术,以下每个章节都旨在帮助你使用 Seedance 3.0 生成令人惊叹的 AI 视频。
Seedance 3.0 提示词是你向 AI 视频引擎提供的文本描述,用于生成视频。你的 Seedance 提示词告诉模型要展示什么主体、创建什么场景、应用哪种视觉风格以及摄像机如何移动。提示词的质量直接决定了生成 AI 视频输出的质量。
Seedance 3.0 使用先进的语言模型来解读自然语言提示词,并将其转换为电影级视频序列。与简单的基于关键词的视频生成器不同,Seedance 3.0 中的提示词引擎能理解上下文、空间关系和时间流——这意味着更详细的提示词能产生更出色的 AI 视频效果。

当你在 Seedance 3.0 视频生成器中输入 Seedance 提示词时,AI 会处理你的文本并根据视频设置生成 5 到 15 秒的视频片段。学会如何撰写有效的提示词是使用 Seedance 3.0 创建专业 AI 视频最重要的技能。
撰写出色的 Seedance 提示词遵循经过验证的四步结构。每个要素都建立在前一个之上,形成一个完整的提示词,从 Seedance 3.0 中持续输出高质量的 AI 视频。
任何有效的 Seedance 提示词的第一个要素是清晰的主体描述。准确告诉 AI 视频生成器视频中应该出现什么或谁。对物理特征、位置和主体数量要具体说明。
弱提示词: "城市中的一个人"
强 Seedance 提示词: "一个穿着红色皮夹克的年轻女性,自信地走过夜晚被雨水浸湿的街道"
提示词中的主体描述越具体,你对视频输出的控制力就越强。在 AI 视频中包含关于服装、表情、肢体语言以及主体与之互动的物体的细节。
在确定主体之后,在 Seedance 提示词中描述周围环境。当提示词为 AI 视频引擎提供足够的空间上下文时,视频生成器能出色地渲染复杂环境。
包含以下细节:
一个好的 Seedance 提示词环境描述能将简单的主体转变为完整的视觉叙事。当你的提示词为主体周围的世界描绘出完整画面时,AI 视频生成器会创作出更加身临其境的视频。
Seedance 3.0 通过你的提示词支持多种视觉风格。这是你的 Seedance 提示词从基本描述进阶到创意视频指导的地方。
在 Seedance 提示词中用于 AI 视频生成的风格关键词包括:
在 Seedance 提示词中添加情绪指示词——如"忧郁"、"活力四射"、"神秘"或"宁静"——有助于 AI 视频引擎匹配生成视频的情感基调。
有效 Seedance 提示词的最后一层是镜头指导。Seedance 3.0 理解专业摄影术语,因此在提示词中包含镜头指令能让你对最终视频输出拥有导演级的控制力。
Seedance 提示词的镜头运动:
在 Seedance 提示词中包含镜头指导是区分业余 AI 视频和 Seedance 3.0 生成的电影级视频的关键。
以下是按视频类别整理的经过测试的 Seedance 提示词示例。每个提示词都遵循上述四步结构精心制作。将这些作为你自己 AI 视频提示词创作的模板,或直接在 Seedance 3.0 中试用。
电影级 Seedance 提示词专注于戏剧性构图、电影级灯光和专业镜头运作。这些提示词能产出可媲美专业视频制作的电影级 AI 视频序列。

"A lone figure standing on the edge of a cliff overlooking a vast misty valley at golden hour. Volumetric sunlight streaming through clouds, lens flare, warm color grading. Slow dolly forward with subtle parallax, cinematic 2.39:1 aspect ratio"
"Two silhouettes facing each other on a rooftop during a thunderstorm, lightning illuminating the scene in brief flashes. Rain droplets caught in slow motion, dramatic backlighting. Orbiting camera movement, film noir aesthetic"
"A vintage train emerging from a tunnel into a snow-covered mountain landscape. Steam billowing from the locomotive, soft winter light, shallow depth of field. Tracking shot following the train from the side, Wes Anderson color palette"
自然风景类 Seedance 提示词利用 AI 视频生成器在每一帧中渲染逼真环境、水动态和自然光照的能力。

"Snow-capped mountain peaks reflected in a crystal-clear alpine lake at sunrise. Morning mist rising gently from the water surface, wildflowers swaying in a light breeze in the foreground. Slow aerial drone shot pulling back to reveal the full panorama, photorealistic"
"A dense tropical rainforest canopy seen from below, with beams of golden sunlight filtering through the leaves. Exotic birds flying between the branches, rain droplets falling in slow motion. Smooth upward crane shot, National Geographic documentary video style"
叙事类 Seedance 提示词创建具有人物主体和情感深度的叙事驱动视频场景。AI 视频生成器能以出色的自然动作和真实视频输出处理角色驱动的提示词。

"A woman pushing open the door of a vintage cafe, warm golden light spilling from inside. She smiles at the barista who is preparing her regular coffee. Close-up on the steaming cup being placed on the wooden counter. Soft bokeh background, indie film color grading, handheld camera feel"
"An elderly artist painting on a canvas in a sunlit studio, surrounded by years of finished works leaning against the walls. Paint-stained hands moving with practiced precision, dust particles floating in light beams. Slow push-in from medium shot to close-up of the brush strokes"
以产品为核心的 Seedance 提示词能产出精美的商业级 AI 视频。Seedance 3.0 非常适合创建营销视频内容、产品视频展示和视频广告概念。

"A pair of sleek black wireless headphones floating and slowly rotating above a reflective dark surface. Subtle smoke effects drifting beneath, gradient background from deep navy to black. Professional studio lighting with soft rim light, 360-degree orbiting camera"

"A luxury wristwatch resting on a marble surface, golden hour light hitting the dial creating subtle reflections. Water droplets condensing on the glass face, extreme macro detail. Slow dolly-in from medium to extreme close-up, premium commercial video photography style"
"A frosted glass perfume bottle on a bed of white silk, soft directional light creating elegant shadows. Rose petals falling in slow motion around the bottle. Smooth 360-degree orbit, luxury fashion video advertisement aesthetic"
科幻类 Seedance 提示词突破 AI 视频生成的创作边界,在每一帧中融入未来感环境、奇幻元素和富有想象力的世界构建。

"A cyberpunk cityscape at night with neon-lit skyscrapers, massive holographic advertisements displaying a woman's face and a glowing cat, flying vehicles streaming between buildings. Rain-slicked streets reflecting colorful lights, people with umbrellas below. Slow descending crane shot from above, Blade Runner aesthetic"
"A vast alien desert planet with two suns setting on the horizon, casting long purple and orange shadows across sand dunes. A lone spacecraft descending with landing thrusters firing, kicking up dust clouds. Wide establishing shot, epic sci-fi scale, Dune-inspired cinematography"
城市类 Seedance 提示词以 AI 视频形式捕捉城市生活的活力和氛围。视频生成器能渲染出细节丰富的城市环境,具有真实纹理和自然的人物运动。

"A person with a backpack walking through a trendy neighborhood street lined with vintage cafes and colorful street art murals. Late afternoon golden hour light casting long shadows, shallow depth of field with bokeh highlights. Tracking shot following from behind, indie film aesthetic"
"A busy Tokyo intersection at night, hundreds of people crossing in different directions under massive illuminated billboards. Reflections on wet pavement, neon signs in Japanese characters. High-angle wide shot, time-lapse feel transitioning to normal speed, Lost in Translation mood"
这些 Seedance 提示技巧专门适用于文本转视频生成。将这些技术与四步提示词结构结合使用,以获得 Seedance 3.0 最佳 AI 视频效果。
有效的 Seedance 提示词依赖精确、描述性的语言,而不是模糊的指示。当你的提示词为视频输出提供具体的视觉细节时,AI 视频生成器的响应效果最好。
不要写: "一个漂亮的日落" 应该写: "温暖的金色日落,粉红和橙色条纹穿过多云的天空,太阳半隐在地平线后面,金色光芒在平静的海面上反射"
用具体的感官细节替换"美丽"或"惊人"等抽象形容词。你的 Seedance 提示词应该像一幅画,让摄影师能够将其重现为真实的视频。你提供的视觉信息越多,AI 视频就越精确和惊艳。
Seedance 3.0 生成不同长度的 AI 视频。你的提示词应该考虑到视频的预期时长。5 秒的提示词应该描述一个聚焦的瞬间,而 10 到 15 秒的 Seedance 提示词可以包含渐进的动作或场景发展。
对于较短的视频提示词(5 秒),专注于一个清晰的动作或揭示。对于较长的 AI 视频,描述一个视频序列:"从一个全景建立镜头开始,然后缓慢推进以揭示……"
镜头指导是任何用于 AI 视频创作的 Seedance 提示词中最具影响力的元素之一。包含特定的摄影术语能大幅提升生成视频的视觉质量。
Seedance 提示词的有效镜头关键词:推轨、横摇、俯仰、起重机、跟踪、静态、手持、航拍、特写、全景、中景、过肩、主观镜头、变焦转移。将镜头指导与速度指示词如"缓慢"、"平滑"或"渐进"结合使用,以获得更可控的 AI 视频输出。
Seedance 图片转视频功能将静态图片转换为动态视频序列。在创建 AI 视频时,为图片转视频写提示词需要与文本转视频不同的方法。
使用图片转视频模式时,你的 Seedance 提示词应该描述你想在视频中看到的运动和变化——而不是图片内容本身,因为 AI 已经能看到上传的图片。

例如,如果你上传了一张人像照片,不要写"一个黑发女性"(这是重复 AI 视频引擎已经看到的内容),而应该写一个专注于视频运动的 Seedance 提示词:
"The woman turns her head slowly to the left, her hair flowing with the movement. A gentle breeze moves through the scene. Soft bokeh background shifts with subtle parallax. Slow motion, cinematic"
这种方法给 AI 视频生成器明确的动画方向,同时尊重源图片中已经存在的视觉信息。
在为图片转视频功能撰写 Seedance 提示词时,遵循以下最佳实践:
为了获得图片转视频的最佳效果,上传高分辨率图片并撰写描述 AI 视频中渐进、可信运动的 Seedance 提示词。
一旦你掌握了基础知识,这些高级提示词技术将帮助你在 Seedance 3.0 中持续获得专业级的 AI 视频效果。
使用这个经过测试的 Seedance 提示词公式作为任何 AI 视频生成的模板:
[主体] + [动作] + [环境] + [灯光/氛围] + [镜头运动] + [风格参考]
使用 Seedance 提示词公式的视频示例:
"A black cat [subject] leaping gracefully from one rooftop to another [action] across a twilight cityscape with distant twinkling lights [environment], warm amber streetlights creating long shadows [lighting], smooth tracking shot following the jump in slow motion [camera], Studio Ghibli animation style [style]"
这个 Seedance 提示词公式确保每次 AI 视频生成都具有视觉清晰度、叙事方向和风格一致性。通过替换每个要素来调整公式以适应任何视频类型或用途。
即使是有经验的用户也会犯这些降低 AI 视频输出质量的 Seedance 提示词错误:
避免这些常见错误将立即改善你的 AI 视频提示词效果。
好的 Seedance 提示词结合了四个要素:清晰的主体描述、详细的环境和场景、明确的视觉风格或氛围,以及特定的镜头运动指令。最好的提示词读起来像剧本中的一个场景——它们为 AI 视频生成器提供足够的细节来产出电影级视频,同时留有创意解读的空间。
理想的 Seedance 提示词长度在 30 到 150 个英文单词之间。AI 视频引擎在提示词足够详细以指导视频生成但又不至于关键信息被稀释时表现最佳。注重质量而非数量——一个结构良好的 50 词 Seedance 提示词比一个散漫的 200 词描述产出更好的视频效果。
Seedance 3.0 使用英语提示词进行 AI 视频创作效果最佳。虽然引擎能理解多种语言,但英语 Seedance 提示词产出最精确和详细的视频效果,因为 AI 训练数据主要基于英语。如果你使用其他语言书写,建议将提示词翻译为英语以获得最佳视频输出。
Seedance 3.0 提供三种 AI 视频模型:Seedance 3.0 Pro 提供最高质量的视频和最忠实的提示词解读,Seedance 3.0 Standard 在视频质量和速度之间取得平衡,Seedance 3.0 Lite 最适合快速视频迭代。对于包含复杂镜头运动的详细 Seedance 提示词,使用 Pro。对于更简单的提示词,Lite 就足够了。比较视频模型并查看价格以找到最适合你工作流程的选项。
不能。文本转视频提示词应该描述完整的视频场景(包括主体外观),而图片转视频 Seedance 提示词应该只关注运动和变化。如果你将文本转视频的提示词用于图片转视频,AI 视频生成器可能会忽略你上传的图片并生成冲突的视频画面。
Seedance 3.0 目前在 AI 视频生成中不使用单独的反向提示词字段。相反,在 Seedance 提示词中描述你想要在视频中看到的内容,而不是想要避免的内容。例如,不要写反向提示词"无模糊",而是写"锐利对焦、清晰细节"来引导视频输出到所需的效果。
要改善 AI 视频提示词效果:使用具体的描述性语言代替模糊的形容词,始终包含镜头运动指导,引用知名的视觉风格或电影,并通过每次调整一个元素来迭代。查看 Seedance 3.0 视频案例获取视频灵感,研究哪些 Seedance 提示词能产出最佳 AI 视频效果。
Seedance 3.0 目前没有内置的视频创作提示词生成器,但本指南提供了你撰写有效 Seedance 提示词所需的所有模板和公式。使用上面的 Seedance 提示词公式作为你的起始模板,并针对任何视频类别进行调整。通过实践,为 AI 视频生成撰写有效提示词将成为你的第二天性。
准备好将这些 Seedance 提示词付诸实践了吗? 立即开始使用 Seedance 3.0 免费创建 AI 视频——无需信用卡。应用本指南中的视频提示技术,感受一个精心撰写的 Seedance 提示词对 AI 视频生成效果的巨大差异。